SELF-HEALING DOT-PRODUCT ENGINE
    32.
    发明申请

    公开(公告)号:US20210225440A1

    公开(公告)日:2021-07-22

    申请号:US17223435

    申请日:2021-04-06

    Abstract: A DPE memristor crossbar array system includes a plurality of partitioned memristor crossbar arrays. Each of the plurality of partitioned memristor crossbar arrays includes a primary memristor crossbar array and a redundant memristor crossbar array. The redundant memristor crossbar array includes values that are mathematically related to values within the primary memristor crossbar array. In addition, the plurality of partitioned memristor crossbar arrays includes a block of shared analog circuits coupled to the plurality of partitioned memristor crossbar arrays. The block of shared analog circuits is to determine a dot product value of voltage values generated by at least one partitioned memristor crossbar array of the plurality of partitioned memristor crossbar arrays.

    Self-healing dot-product engine
    33.
    发明授权

    公开(公告)号:US10984860B2

    公开(公告)日:2021-04-20

    申请号:US16364717

    申请日:2019-03-26

    Abstract: A DPE memristor crossbar array system includes a plurality of partitioned memristor crossbar arrays. Each of the plurality of partitioned memristor crossbar arrays includes a primary memristor crossbar array and a redundant memristor crossbar array. The redundant memristor crossbar array includes values that are mathematically related to values within the primary memristor crossbar array. In addition, the plurality of partitioned memristor crossbar arrays includes a block of shared analog circuits coupled to the plurality of partitioned memristor crossbar arrays. The block of shared analog circuits is to determine a dot product value of voltage values generated by at least one partitioned memristor crossbar array of the plurality of partitioned memristor crossbar arrays.

    Iterative write sequence interrupt
    34.
    发明授权

    公开(公告)号:US10671291B2

    公开(公告)日:2020-06-02

    申请号:US15770753

    申请日:2015-11-17

    Abstract: Example implementations relate to memory read requests. For example, an implementation may include tracking progress of an iterative write sequence to write data to a memory element of a memory module. A received read request is detected to be addressed to a memory bank that includes the memory element undergoing the iterative write sequence. Based on the tracked progress, a time is determined to interrupt the iterative write sequence with insertion of the read request. The time aligns between operations of the iterative write sequence and data is returned within a predetermined read latency.

    Tunable and dynamically adjustable error correction for memristor crossbars

    公开(公告)号:US10452472B1

    公开(公告)日:2019-10-22

    申请号:US15997030

    申请日:2018-06-04

    Abstract: A dot-product engine (DPE) implemented on an integrated circuit as a crossbar array (CA) includes memory elements comprising a memristor and a transistor in series. A crossbar with N rows, M columns may have N×M memory elements. A vector input for N voltage inputs to the CA and a vector output for M voltage outputs from the CA. An analog-to-digital converter (ADC) and/or a digital-to-analog converter (DAC) may be coupled to each input/output register. Values representing a first matrix may be stored in the CA. Voltages/currents representing a second matrix may be applied to the crossbar. Ohm's Law and Kirchoff's Law may be used to determine values representing the dot-product as read from the crossbar. A portion of the crossbar may perform Error-correcting Codes (ECC) concurrently with calculating the dot-product results. ECC codes may be used to only indicate detection of errors, or for both detection and correction of results.

    STAGING WRITE REQUESTS
    38.
    发明申请

    公开(公告)号:US20170308296A1

    公开(公告)日:2017-10-26

    申请号:US15136827

    申请日:2016-04-22

    CPC classification number: G06F3/061 G06F3/0659 G06F3/0679 G06F3/0685

    Abstract: In various examples, a memory may comprise a first subarray having an associated first staging buffer, a second subarray having an associated second staging buffer, and request logic. The request logic may: receive a first write request associated with the first subarray, receive a second write request associated with the second subarray, store the first write request in the first staging buffer, store the second write request in the second staging buffer, and execute the first write request and the second write request.

Patent Agency Ranking